NLY: The Dividend Is Covered. What Protects the Capital?

The 60-second read

Five things to know

1. Coverage improved, but capital fell. EAD of $0.80 exceeded the $0.75 dividend. Book value fell from $20.15 to $17.86 and economic return was negative 7.6%.
2. Reported margin and underlying spread moved differently. GAAP net interest margin increased to 1.54%, but economic margin excluding the premium-amortization adjustment fell to 1.69%.
3. Growth in total earnings is not the same as growth per owner. Common EAD rose 27.0% year over year; per-share EAD rose 9.6% and ending shares rose 15.2%.
4. Mortgage investments are not factory capex. Vendor capital expenditure includes MSR purchases. We use actual cash-flow labels and common-equity valuation rather than a misleading industrial free-cash-flow model.
5. The quote predates the print. $18.29 is the October 20 regular close, not a post-earnings return. The release was after the close; its exact minute is unconfirmed.

Three scenarios, one probability-weighted number

Analyst equity scenarios, not issuer guidance or guaranteed distribution levels.

Signposts: what would change our mind

These are our prospective tests, not management promises. $12.75 is a 25% margin of safety below the $17 base; a lower price is insufficient if funding, book value or dividend capacity deteriorates.

The tape

The valuation reference is the saved October 20, 2026 regular-session close of $18.29. Its provider timestamp is 15:00:02 CDT, before Annaly’s after-close earnings release. The 3.2% regular-session gain therefore cannot be described as the market’s response to these results. A later saved indicative bid of $18.22 and ask of $18.43, timestamped 17:36:09 CDT, does not establish an executed price. We preserve that observation separately and do not quietly substitute its midpoint into the model.

The historical price series supplies context, not a forecast. Our $17 central value and the quarter-end $17.86 common book value are different reference points. Book is a marked balance-sheet measure, while our value includes a required return for taking leveraged mortgage exposure. Neither line is a floor under the stock. The $17.70–$24.52 saved 52-week range also does not establish what the next year must look like.

The vendor’s $13.785 billion market capitalization implies about 753.7 million shares at $18.29, whereas the current issuer reports 784.673 million ending common shares. That is a denominator mismatch, not a reason to alter the filing. We retain the observed price and use primary common book and share counts. We also avoid labelling a subsequent-close window an isolated earnings effect: rates, mortgage spreads and other market news can move the stock on the same day.

The print

The central tension is visible without an adjusted metric debate. Annaly reported a $1.001 billion GAAP net loss and a $1.043 billion loss available to common shareholders, or negative $1.36 per basic share. It also reported $649.140 million of total earnings available for distribution, with $607.835 million attributable to common shareholders and $0.80 per average common share. The latter covered the $0.75 declared dividend. They answer different questions and should remain side by side. Q3 earnings release supplies both the GAAP statement and the full non-GAAP reconciliation.

Common EAD increased 27.0% from the year-earlier $478.755 million, while EAD per share increased 9.6% from $0.73. The difference is meaningful: capital raised and new shares can grow the pool of earning assets without producing the same growth for each existing owner. Ending shares increased 15.2% year over year. This is not proof that every issuance destroyed value; management describes its issuance as accretive. It is a reason to evaluate the per-share result separately from the total.

Quarter-end book value fell 11.4% to $17.86 from $20.15. Adding the $0.75 dividend to that book change gives a computed negative 7.64% economic return, matching the issuer’s rounded negative 7.6%. A covered distribution can coexist with a negative economic result. The prior quarter’s $0.79 EAD and $0.75 dividend already showed coverage, so the latest penny of EAD improvement does not by itself reverse the capital story.

Portfolio and business model

Annaly is a mortgage investment company, not a property landlord collecting rent or an industrial company selling products. Its investment groups are Agency, Residential Credit and mortgage servicing rights, or MSR. The economic portfolio is $106.5 billion: $91.8 billion Agency, $10.4 billion Residential Credit and $4.3 billion MSR. Those are portfolio values, not quarterly revenues. Allocated capital is 55%, 23% and 22%, respectively; the smaller MSR asset pool can consume a much larger share of capital because leverage differs across strategies.

Agency exposure provides mortgage cash flows with agency credit support, but that does not remove price, prepayment, funding or basis risk. Residential Credit introduces borrower and collateral performance as well as financing exposure. MSR earns servicing economics and can help offset some rate sensitivity, yet its valuation, amortization and customer prepayment patterns also matter. Diversification should be judged from aggregate economic results, not from the mere existence of three strategy names.

The $146.147 billion consolidated asset balance is larger than the economic portfolio because consolidation and financing definitions differ. Securitization vehicles bring assets and nonrecourse liabilities onto the GAAP balance sheet. TBA exposure and retained investments introduce further presentation differences. We do not subtract all $127.231 billion of GAAP debt from the $106.5 billion economic portfolio and call the result common value. That would combine unlike denominators. The company’s common equity and its reconciliations are the appropriate starting point.

Annaly reports three operating segments, with Corporate and Other reconciling to consolidated results. The operating committee uses comprehensive income as its segment profit measure. We have reconciled eight historical quarters from actual segment tables, deriving fourth-quarter flows as annual less nine-month results. The Q3 2026 segment comprehensive-income table is not yet furnished in the released materials. We show the available history and explicitly leave that current cell unavailable, rather than inventing segment profit from capital percentages. Q3 financial supplement and Q3 investor presentation supply the current strategy portfolio mix.

How carry and marks diverged

Interest income was $1,970.425 million and interest expense $1,426.813 million, leaving $543.612 million of GAAP net interest income. Net servicing income added $156.009 million. Those positive flows did not prevent a large GAAP loss: investment losses of $4,721.745 million were only partly offset by derivative gains of $3,074.072 million. The net of those two lines was negative $1,647.673 million. A hedge gain can be substantial and still fail to preserve common capital when the assets and hedges do not move together.

The margin comparison needs the same care. GAAP net interest margin rose from 1.47% to 1.54%, but economic net interest margin excluding the premium-amortization adjustment declined from 1.76% to 1.69%. Economic spread fell eight basis points to 1.42%, as economic funding cost increased from 3.96% to 4.03%. The premium-amortization adjustment changed from a $7.081 million cost to a $44.518 million benefit, a favorable $51.599 million swing. That accounting movement is a reason to retain the ex-adjustment view.

The common book bridge makes the capital result tangible. Starting at $20.15, subtract $1.36 GAAP earnings loss per share, $0.31 other comprehensive loss and the $0.75 dividend, then add $0.13 of other effects to reach $17.86. The other line should not be relabelled entirely as ATM issuance accretion: it includes the issuer’s other bridge effects. Distributions reduce book when paid out even when they are funded by recurring earning capacity. That is why dividend yield and book return must be read together.

Earnings quality

The EAD reconciliation removes large valuation effects; that is useful for estimating recurring carry, but it is not a declaration that marks are economically irrelevant. Starting with the $1,000.698 million GAAP net loss, the release adds back $4,728.272 million of investment losses, subtracts $2,982.293 million of derivative gains and makes the additional items shown below to reach $649.140 million total EAD. Preferred dividends of $41.305 million then leave $607.835 million common EAD. Small differences from the headline GAAP mark lines reflect the exact adjustments; do not substitute one set for the other.

Two adjustments deserve particular attention. MSR amortization of $93.570 million is deducted in the EAD reconciliation, so the $156.009 million net servicing line is not equivalent to distributable servicing cash. Their difference is $62.439 million before other relevant items. The $10.738 million securitization-cost add-back is labelled nonrecurring by the issuer, but repeated securitization activity means investors should watch the cost pattern. We do not assume that every future funding transaction becomes costless.

The $0.05 per-share gap between EAD and the dividend is helpful but thin. Dividing current common EAD by ending shares gives $0.775 rather than reported $0.80, or about 1.03 times the dividend. This is only a denominator sensitivity: shares were issued during the quarter, so it is not an alternative reported EPS or an exact cash-distribution coverage calculation. It illustrates how new capital must keep earning after the shares enter the full-period denominator.

The standard quality checklist is adapted to the business rather than populated with false industrial metrics. There is no inventory-days thesis and no trade-receivables DSO inference for a mortgage portfolio. Current-quarter operating cash flow has not been published in the earnings materials; the latest filed quarterly cash-flow statement is for June. EAD, GAAP income and taxable income differ, and REIT distribution rules do not turn an EAD payout ratio into a guarantee.

Funding and capital allocation

Cash was $2.181 billion, while reported available liquidity was $8.5 billion. The broader amount includes $6.9 billion of unencumbered assets and $1.6 billion of collateral pledged for future advances. Cash plus unencumbered Agency assets was $4.2 billion. These are related but not interchangeable measures. Warehouse capacity of $8.3 billion, including $2.8 billion committed, is access to funding rather than cash already owned. The liquidity cushion matters because collateral demands can rise just as market values fall.

Economic leverage increased to 5.9 times from 5.6 times and GAAP leverage was 8.0 times. Recourse debt of $86.894 billion differs from economic debt of $94.008 billion and consolidated debt of $127.231 billion. Each denominator serves a purpose; using them interchangeably would obscure rather than clarify the financing risk. The current weighted-average repo maturity was 30 days. Funding therefore has to be rolled repeatedly, and haircuts or counterparty appetite can change faster than mortgage assets mature.

Historical cash statements reinforce why a generic free-cash-flow screen is unreliable here. Operating cash was $692.911 million in 2025, compared with $3,310.659 million in 2024 and $2,367.203 million in 2023. Mortgage investment transactions and accounting classifications make those figures different from a manufacturer’s operating cash. The provider’s 2025 capex of $915.074 million equals MSR purchases of $882.834 million plus $32.240 million of interests in MSR. We reject the industrial capex label and do not subtract it to advertise a conventional FCF yield.

Annaly raised $757 million of net ATM common capital during the quarter. Issuance can expand earning assets and be accretive when priced advantageously, but ownership per share and risk exposure remain the tests. Separately, 17.7 million Series I preferred shares were redeemed for $442.5 million on October 1. The related delisting concerned that preferred series, not NLY common. September 30 still includes the Series I balance, so we do not manufacture a quarter-end common-equity gain from the subsequent redemption.

The company’s common repurchase authorization is $1.5 billion through December 31, 2029. The annual reports state no program repurchases in the reviewed years. Withholding on stock awards is a separate cash-flow item, not evidence of an open-market capital-return program. Dividends paid in the annual statement include all classes; they should not be compared with a common-only numerator without reconciliation. These distinctions are more useful than a headline claim that all financing cash has reached existing common owners.

Common-equity valuation

Our call is HOLD at 2/5 conviction and High uncertainty. Central value is $17 per common share, with a $13–$22 scenario range and $17.25 weighted value. At the dated $18.29 close, the central price return is negative 7.1%. If a full $3 annual dividend were maintained and the price reached $17 one year later, the simple total return would be 9.3%. That conditional return is positive, but below our 16.5% required return in the dividend model. A high cash yield is not automatically sufficient compensation.

Route one discounts common dividends of $2.80, $2.80, $2.85, $2.90 and $2.95 over five years at a 16.5% required equity return, with zero terminal growth. The explicit dividends have present value $9.218. Terminal value is $17.879 at year five, with present value $8.331, for total value $17.549. The distribution path is our analytical normalization below the current $3 run rate, not management guidance or a forecast of an announced cut. Zero long-run growth avoids assuming that leveraged asset growth automatically compounds value for each share.

Route two applies 0.95 times the $17.86 marked common book value, giving $16.967. The modest discount reflects liquidity, spread and funding risks rather than a liquidation forecast. Route three assumes sustainable common return on equity of 15% and a required return of 16%, with zero growth: $17.86 times 15% divided by 16% equals $16.744. This is a justified-book-value cross-check. The three values are correlated because they depend on common earning power and capital preservation; their clustering does not remove model risk.

These are equity models. Dividends and common book are already subordinate to debt and preferred claims. Adding cash or subtracting all repo and securitization debt again would double count the financing structure. We therefore use no enterprise-value-to-equity bridge, WACC or industrial free-cash-flow multiple. There is also no unsupported recovery value for a historic acquisition or preferred redemption. A mortgage REIT can distribute cash while capital shrinks, so sustainable distribution capacity and book resilience must both support the price.

The reverse calculation is simple and useful: an $18.29 price at a 16.5% required return implies a perpetual annual common dividend of $3.018, slightly above the current $3.00 run rate. This is a level-perpetuity illustration, not the five-year normalized path. At a 16% required return, the price-to-book ratio implies a sustainable ROE of 16.4%. Both calculations ask whether today’s carry and capital can persist without additional risk or dilution. They are expectations implied by chosen assumptions, not a survey of investor beliefs.

The sensitivity grid separately values constant annual dividends at alternative required returns. Its center, $2.80 divided by 16.5%, is $16.97; it is deliberately not labelled the $17.55 five-year model. The bear and bull values use $2.40/18.5% and $3.30/15%, respectively. We require a 25% margin below the $17 central value before considering a price-driven upgrade, or $12.75, and even then the funding and operating facts must remain intact. A falling quote cannot compensate for an unbounded deterioration in capital.

Wall Street context

The saved Street record contains 16 Buy, 10 Hold and 2 Sell ratings, with an average target of $22 and a $21–$23 target range. Our $17 central value is lower and our HOLD is more cautious than that favorable aggregate tilt. These are saved provider observations, not evidence that every analyst has incorporated tonight’s release. The most recent retained target item is BTIG’s October 6 move to $21 from $24. An older UBS item raised its target to $23 from $22.50 on April 24. Neither is a verified post-print revision.

Saved annual EPS estimates are approximately $3.10 for 2026, $3.13 for 2027 and $3.03 for 2028, with seven, seven and one contributing analysts in the file. The provider definition is not established as GAAP, so these numbers must not be compared directly with the quarter’s negative $1.36 GAAP EPS. The thin 2028 coverage is especially important. The saved revenue estimate is also not reconciled to gross interest income or net interest income; we do not build a revenue growth thesis from that ambiguous field.

The strongest bullish interpretation is that the current dividend is covered, a large liquidity pool provides flexibility, and stabilized mortgage spreads could improve book while funding conditions support carry. The counterargument is that the current quarter already demonstrates negative economic return despite coverage, rising leverage and narrower underlying spread. A target near $22 requires the favorable path to persist. Our disagreement concerns the sustainable economics and required compensation for risk, not a claim that the Street has made an accounting error.

Management and execution

Annaly is internally managed. The annual reports list 212 employees at year-end 2025, compared with 191 and 187 in the preceding years, and Ernst & Young LLP as auditor. Its strategy requires funding access, investment selection, hedging and capital allocation to work together. A larger portfolio or a larger employee base does not prove greater value per share. We assess management against current distribution capacity, economic return, funding resilience and the economics of common issuance rather than an industrial sales-growth target.

The quarter earns credit for maintaining EAD coverage and available liquidity. It earns less credit for capital preservation: book fell 11.4%, economic return was negative 7.6%, and economic leverage rose. The 100% hedge ratio should be interpreted within the issuer’s definition, not as full insurance against every rate, spread and prepayment outcome. The GAAP mark bridge shows why a portfolio can be heavily hedged and still lose capital.

There is no verified earnings-call commentary yet. The announced call is October 21 at 08:00 CDT, which is separate from the October 20 after-close release. We do not turn that scheduled call time into a release timestamp or invent management answers. The next questions are specific: how much of current EAD survives the larger share denominator, how funding cost and asset yields change, what issuance contributes to book per share, and how liquidity performs under stressed haircuts.

No earlier exact-period episode is being overwritten, and no prior Charged Alpha NLY investment angle was verified in the current canonical source check. This is an initial source-backed assessment in this workflow, not an invented upgrade or downgrade. Future changes should be explained against these saved assumptions and checkpoints so a rising dividend does not erase a deteriorating capital record, or a weak quarter obscure genuine subsequent improvement.

Ranked risks and counterarguments

First is mortgage-spread risk. The issuer’s separate instantaneous shock analysis estimates a 10.1% NAV decline for a 25-basis-point widening in MBS spreads, versus a 10.4% increase for a 25-basis-point tightening. Second is funding and liquidity risk: repo rolls frequently, collateral values can fall and counterparties can demand additional margin. The annual filing explains that forced sales can occur when conditions are unfavorable. Available liquidity reduces that risk but does not make leverage harmless.

Third is rate and prepayment risk. The issuer estimates a 5.8% NAV decline for a 75-basis-point upward rate shock and a 3.4% increase for the opposite shock. Experienced CPR was 9.6%, down from 11.6%, but future prepayment behavior affects premiums, reinvestment and MSR value. These shocks are separate, static and assume no active management. We do not add rate and spread scenarios together or translate NAV changes mechanically into guaranteed stock-price returns.

Fourth is distribution and denominator risk. Current per-share coverage is modest, and newly issued common shares must earn enough to sustain the full-period payout. Fifth is credit, model and servicing risk within Residential Credit and MSR, including counterparty performance and securitization execution. Sixth is regulatory and tax structure. REIT qualification generally requires at least 90% of taxable income to be distributed, while the company generally aims to distribute all taxable income. That is not a guarantee of a particular common dividend and does not equate taxable income with EAD.

The charter’s 9.8% ownership restriction supports the REIT structure but is not evidence of a controlling shareholder. Historical MTGE acquisition filings describe a completed transaction, not a currently pending merger that changes the covered common security. The October preferred redemption is likewise a current capital-allocation event, not a common-stock delisting. These distinctions preserve valid coverage while keeping the actual financial risks visible. The current source review found no basis to exclude the continuing common issuer on those historical or series-specific events.

Decision checkpoints

The immediate information event is the October 21 earnings call, but the quality of the answers matters more than the occurrence of the event. We want a clear explanation of mortgage-spread performance, funding costs, capital raised and the path from total EAD to per-share earning capacity. Subsequent book-value and portfolio disclosures can test the assumptions before a full year passes. We will distinguish a marked recovery from genuinely stronger recurring earning power rather than call either one sufficient by itself.

The next quarterly release should show whether economic spread can recover to at least 1.50% without taking economic leverage above our 6.0-times green threshold. The dividend test is at least 1.10-times per-share coverage for two quarters, not just one favorable penny. Those are analytical thresholds, not company guidance. They deliberately ask for a buffer because the distribution and financing structure leave common shareholders exposed to adverse capital movements.

The investment conclusion can change in either direction. Better spread, stable book and sustained coverage would strengthen the case for the $22 bull scenario. Deteriorating funding, a shrinking capital base or distributions that exceed sustainable capacity would move the evidence toward $13 or lower. At $18.29 the stock offers conditional income but insufficient margin for a higher-conviction purchase under our assumptions. HOLD expresses that balance; it is not a promise that capital or the dividend will be stable.

1. What recurring earning capacity remains after the new shares enter a full-quarter denominator?
2. Does the next book update stabilize at or above $17.86 while the dividend is earned?
3. Can economic spread recover without rising leverage or weaker liquidity?
4. What funding and collateral assumptions explain management’s stress resilience?
5. Would a lower price coincide with intact facts, or merely reflect deteriorating common capital?

Sources and methods

This packet uses the actual October 20 Q3 earnings release, financial supplement and investor presentation, nine quarterly earnings releases, three audited annual reports, the latest available June 30 Form 10-Q, and saved FMP market and Street data. Every downloaded body and extracted table is retained with a SHA-256 binding in the local source inventory. The source hierarchy matters: the issuer defines reported numbers and accounting categories; the market provider supplies dated prices and expectations; our model supplies explicitly labelled assumptions.

The current Q3 10-Q and cash-flow statement were not available in the reviewed release set. That is a scope limit, not proof of a filing delay. We do not invent Q3 operating cash, a cash-conversion ratio or a tax reconciliation. Annual comparative cash statements preserve exact labels. In particular, a standardized SEC dividend tag mapped to a declared-but-unpaid supplemental amount, and provider capex mapped to MSR investment. Both were rejected as substitutes for the primary table meanings.

Values are USD millions unless labelled billions, per share, percent, basis points or a ratio. Annualized dividends multiply the current quarter by four and are conditional. Economic return uses book change plus the declared dividend over starting book; price return and total return use separately stated market-price assumptions. Average-share EAD and ending-share sensitivities are never treated as identical. Rounded display figures can differ slightly from calculations using full precision in the saved dataset.

The valuation is common-equity analysis with no duplicate debt or cash adjustment. Scenario probabilities are subjective and sum to 100%. Our required return, normalized distribution path and book discount are judgments, not issuer forecasts. Historical returns, static issuer shock models and a saved Street consensus cannot guarantee a future outcome. The analysis is an initial assessment of this print, and the exact release minute remains unconfirmed rather than replaced with an SEC acceptance time or scheduled call time.

The financial model and source dataset are retained with the episode, including provider discrepancies, unreconciled fields and the full GAAP-to-EAD bridge. No quarterly release figure is inferred from a calendar entry. No post-print analyst revision or conference-call answer is asserted without an actual source. This makes later reassessment possible without rewriting what was known at the time of the initial decision.